---
license: apache-2.0
dataset_info:
- config_name: filtered
  features:
  - name: ID
    dtype: int64
  - name: CITY
    dtype: int64
  - name: COUN
    dtype: int64
  - name: SUBJ
    dtype: int64
  - name: SEX
    dtype: int64
  - name: AGE
    dtype: int64
  - name: RELI
    dtype: int64
  - name: PRAC
    dtype: int64
  - name: FOCC
    dtype: int64
  - name: MOCC
    dtype: int64
  - name: FIEL
    dtype: int64
  - name: EMOT
    dtype: int64
  - name: WHEN
    dtype: int64
  - name: LONG
    dtype: int64
  - name: INTS
    dtype: int64
  - name: ERGO
    dtype: int64
  - name: TROPHO
    dtype: int64
  - name: TEMPER
    dtype: int64
  - name: EXPRES
    dtype: int64
  - name: MOVE
    dtype: int64
  - name: EXP1
    dtype: int64
  - name: EXP2
    dtype: int64
  - name: EXP10
    dtype: int64
  - name: PARAL
    dtype: int64
  - name: CON
    dtype: int64
  - name: EXPC
    dtype: int64
  - name: PLEA
    dtype: int64
  - name: PLAN
    dtype: int64
  - name: FAIR
    dtype: int64
  - name: CAUS
    dtype: int64
  - name: COPING
    dtype: int64
  - name: MORL
    dtype: int64
  - name: SELF
    dtype: int64
  - name: RELA
    dtype: int64
  - name: VERBAL
    dtype: int64
  - name: NEUTRO
    dtype: int64
  - name: Field1
    dtype: string
  - name: Field3
    dtype: int64
  - name: Field2
    dtype: int64
  - name: MYKEY
    dtype: int64
  - name: SIT
    dtype: string
  - name: STATE
    dtype: int64
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: train
    num_bytes: 2688089
    num_examples: 5874
  - name: validation
    num_bytes: 335896
    num_examples: 734
  - name: test
    num_bytes: 336354
    num_examples: 735
  download_size: 799248
  dataset_size: 3360339
- config_name: original
  features:
  - name: ID
    dtype: int64
  - name: CITY
    dtype: int64
  - name: COUN
    dtype: int64
  - name: SUBJ
    dtype: int64
  - name: SEX
    dtype: int64
  - name: AGE
    dtype: int64
  - name: RELI
    dtype: int64
  - name: PRAC
    dtype: int64
  - name: FOCC
    dtype: int64
  - name: MOCC
    dtype: int64
  - name: FIEL
    dtype: int64
  - name: EMOT
    dtype: int64
  - name: WHEN
    dtype: int64
  - name: LONG
    dtype: int64
  - name: INTS
    dtype: int64
  - name: ERGO
    dtype: int64
  - name: TROPHO
    dtype: int64
  - name: TEMPER
    dtype: int64
  - name: EXPRES
    dtype: int64
  - name: MOVE
    dtype: int64
  - name: EXP1
    dtype: int64
  - name: EXP2
    dtype: int64
  - name: EXP10
    dtype: int64
  - name: PARAL
    dtype: int64
  - name: CON
    dtype: int64
  - name: EXPC
    dtype: int64
  - name: PLEA
    dtype: int64
  - name: PLAN
    dtype: int64
  - name: FAIR
    dtype: int64
  - name: CAUS
    dtype: int64
  - name: COPING
    dtype: int64
  - name: MORL
    dtype: int64
  - name: SELF
    dtype: int64
  - name: RELA
    dtype: int64
  - name: VERBAL
    dtype: int64
  - name: NEUTRO
    dtype: int64
  - name: Field1
    dtype: string
  - name: Field3
    dtype: int64
  - name: Field2
    dtype: int64
  - name: MYKEY
    dtype: int64
  - name: SIT
    dtype: string
  - name: STATE
    dtype: int64
  splits:
  - name: train
    num_bytes: 2731655
    num_examples: 6132
  - name: validation
    num_bytes: 341234
    num_examples: 766
  - name: test
    num_bytes: 342125
    num_examples: 768
  download_size: 766145
  dataset_size: 3415014
configs:
- config_name: filtered
  data_files:
  - split: train
    path: filtered/train-*
  - split: validation
    path: filtered/validation-*
  - split: test
    path: filtered/test-*
- config_name: original
  data_files:
  - split: train
    path: original/train-*
  - split: validation
    path: original/validation-*
  - split: test
    path: original/test-*
task_categories:
- text-classification
language:
- en
tags:
- medical
pretty_name: ISEAR Dataset - International Survey on Emotion Antecedents and Reactions
size_categories:
- 1K<n<10K
---

# ISEAR Dataset - International Survey on Emotion Antecedents and Reactions

## Dataset Summary

The ISEAR (International Survey on Emotion Antecedents and Reactions) dataset contains 7,666 personal accounts of emotional experiences collected by psychologists Klaus R. Scherer and Harald G. Wallbott as part of a large-scale international study on emotions. Respondents from 37 countries were asked to describe situations where they experienced specific emotions (joy, fear, anger, sadness, disgust, shame, or guilt), along with detailed information about their reactions and appraisals.

This dataset is a cleaned version of the original ISEAR questionnaire data, prepared for natural language processing and emotion classification tasks.

## Dataset Structure

The dataset is provided in two configurations:

- `original`: The complete dataset with basic cleaning
- `filtered`: A subset that removes very short descriptions (under 25 characters)

### Data Instances

Each instance in the dataset represents one emotional experience and contains:

- A textual description of the situation that elicited the emotion
- The emotion label (joy, fear, anger, sadness, disgust, shame, or guilt)
- Detailed metadata about physiological responses, cognitive appraisals, and demographic information

Example instance:

```
{
  "SIT": "When I got a letter offering me the Summer job that I had applied for.",
  "EMOT": 1,  # Joy
  "WHEN": 3,  # months ago
  "LONG": 2,  # an hour
  "INTS": 2,  # intense
  "ERGO": 0,  # ergotropic arousal score
  "TROPHO": 0,  # trophotropic arousal score
  "TEMPER": 1,  # felt temperature score
  ...
}
```

### Data Fields

#### Main Fields for Classification Tasks

- `SIT`: Textual description of the situation that elicited the emotion
- `EMOT`: Emotion category (1=joy, 2=fear, 3=anger, 4=sadness, 5=disgust, 6=shame, 7=guilt)

#### Temporal and Intensity Information

- `WHEN`: When the situation happened (1=days ago, 2=weeks ago, 3=months ago, 4=years ago)
- `LONG`: Duration of the emotion (1=minutes, 2=an hour, 3=several hours, 4=a day or more)
- `INTS`: Intensity of the emotion (1=not very, 2=moderately intense, 3=intense, 4=very intense)

#### Physiological and Behavioral Responses

- `ERGO`: Ergotropic arousal score (0-4), including change in breathing, heart beating faster, muscles tensing/trembling, perspiring/moist hands
- `TROPHO`: Trophotropic arousal score (0-3), including lump in throat, stomach troubles, crying/sobbing
- `TEMPER`: Felt temperature (-1 to +2), feeling cold/shivering, feeling warm/pleasant, feeling hot/cheeks burning (0 being assigned when no temperature symptom was mentioned)
- `MOVE`: Movement behavior (-1=withdrawing, 0=no movement, +1=moving towards people and things)
- `EXPRES`: Nonverbal activity score (0-6), including laughing/smiling, crying/sobbing, other facial expression change, screaming/yelling, other voice changes, changes in gesturing
- `PARAL`: Paralinguistic activity score (0-3), speech-melody change, speech disturbances, speech tempo change
- `VERBAL`: Verbal activity (0=silence, 1=short utterance, 2=one/two sentences, 3=lengthy)
- `EXP1`: Laughing/smiling (0=no, 1=yes)
- `EXP2`: Crying/sobbing (0=no, 1=yes)
- `EXP10`: Aggression, moving against people and things (0=no, 1=yes)

#### Cognitive Appraisals

- `CON`: Control/hiding of feelings (1=not at all, 2=a little, 3=very much, 0=not applicable)
- `EXPC`: Expectedness of event (1=not at all, 2=a little, 3=very much, 0=not applicable)
- `PLEA`: Event pleasantness (1=pleasant, 2=neutral, 3=unpleasant, 0=not applicable)
- `PLAN`: Effect on goals/plans (1=helped, 2=didn't matter, 3=hindered, 0=not applicable)
- `FAIR`: Perceived fairness (1=not at all, 2=a little, 3=very much, 0=not applicable)
- `CAUS`: Responsibility attribution (1=self, 2=close persons, 3=other persons, 4=impersonal agency, 0=not applicable)
- `COPING`: Coping strategy (1=I did not think that any action was necessary, 2=I believed that I could positively influence the event and change the consequences, 3=I believed that I could escape from the situation or avoid negative consequences, 4=I pretended that nothing important had happened and tried to think of something else, 5=I saw myself as powerless and dominated by the event and its consequences)
- `MORL`: Perceived as improper or immoral by acquaintances (1=not at all, 2=a little, 3=very much, 0=not applicable)
- `SELF`: Effect on self-esteem (1=negative, 2=none, 3=positive, 0=not applicable)
- `RELA`: Effect on relationships (1=negative, 2=none, 3=positive, 0=not applicable)

#### Demographic Information

- `SEX`: Gender (1=male, 2=female)
- `AGE`: Age of the respondent (numerical value)
- `RELI`: Religion (1=protestant, 2=catholic, 3=Jewish, 4=Hindu, 5=Buddhist, 6=native, 7=others, 8=areligious)
- `PRAC`: Religious practice (1=practicing, 2=not practicing)
- `FOCC`: Father's occupation (1=housewife, 2=unemployed, 3=student, 4=blue collar untrained, 5=blue collar trained, 6=white collar nonacademic, 7=white collar academic, 8=self-employed nonacademic, 9=self-employed academic)
- `MOCC`: Mother's occupation (1-9, same as FOCC)
- `FIEL`: Field of study (1=psychology, 2=social sciences, 3=languages, 4=fine arts, 5=law, 6=natural science, 7=engineering, 8=medical, 9=other)
- `COUN`: Country (1=Sweden, 2=Norway, 4=Finland, 8=Austria, 9=Australia, 10=Brazil, 12=Bulgaria, 16=New Zealand, 17=Netherlands, 19=Spain, 20=Zambia, 22=USA, 26=India, 27=Malawi, 30=China Mainland, 33=Honduras)
- `CITY`: City classification (undocumented in original dataset, appears to have two values, likely indicating urban/rural or size classification)
- `SUBJ`: Subject identifier (undocumented in original dataset, likely a unique ID for each participant)
- `NEUTRO`: Not documented in original materials

#### Note on Demographic Variables

This dataset contains demographic variables that reflect the categorization approaches of when the data was collected (1980s-90s). Some variables like gender (labeled as "SEX" with binary options) and religion do not reflect contemporary understanding of these attributes or modern demographic collection practices. These variables are preserved for research integrity and historical accuracy.

Researchers using this dataset should be mindful when reporting results based on these demographic variables and consider contemporary ethical standards in their analyses.

### Data Splits

The dataset provides standard train/test/validation splits for both configurations.

## Dataset Creation

### Source Data

The original data was collected through questionnaires administered to university students in 37 countries between the 1980s and early 1990s by Klaus R. Scherer (University of Geneva) and Harald G. Wallbott (University of Salzburg). Respondents were asked to recall experiences of seven fundamental emotions and answer detailed questions about their reactions and appraisals.

Official source: [Geneva Emotion Research Group](https://www.unige.ch/cisa/research/materials-and-online-research/research-material/)

### Preprocessing

The following preprocessing steps were applied to create this version:

1. Basic text cleaning (removing unusual line breaks, square brackets, etc.)
2. Conversion of floating-point values to appropriate integers for categorical variables
3. For the `filtered` configuration: removal of very short descriptions (under 25 characters)

## Ethical Usage Guidelines

While this dataset has many potential applications, users should consider these ethical guidelines:

### Recommended Applications
- Academic research on emotional expression and recognition
- Healthcare applications for mental health support under appropriate supervision
- Enhancing communication effectiveness through better emotional awareness
- Supporting team coaching programs that foster psychological safety, enchance communication effectiventess and collaboration

### Usage Considerations
- Some jurisdictions have regulations around emotion recognition technology, particularly in workplace and educational settings
- Privacy and consent are important when analyzing personal emotional expressions

Users should be aware of restrictions on emotion recognition systems and ensure any applications developed using this dataset comply with regulatory requirements. This dataset should not be used for commercial emotion categorization systems to infer emotions of a natural person.

### Discussion of Biases

- **Sample bias**: The original data was collected primarily from university students, which may not represent the general population
- **Cultural representation**: While 37 countries are included in the original study, out of which 16 countries were included in the dataset, representation varies by country
- **Historical context**: Data was collected in the 1980s-90s, and emotional expressions may have cultural changes since then

## Additional Information

### Dataset Curators

The original dataset was created by:

- Klaus R. Scherer, University of Geneva, Switzerland
- Harald G. Wallbott, University of Salzburg, Austria

### Licensing Information

This processed version of the ISEAR dataset is shared under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). The original ISEAR data was collected by the Geneva Emotion Research Group for research purposes.

### Citation Information

If you use this dataset, please cite both the original work and this processed version:

**Original Dataset:**

```
Scherer, K. R., & Wallbott, H. G. (1994). Evidence for universality and cultural variation of differential emotion response patterning. Journal of Personality and Social Psychology, 66(2), 310–328.
```

**Processed Dataset:**

```
Savalera. (2025). ISEAR-NLP: A processed version of the ISEAR emotion dataset. Hugging Face Datasets. https://huggingface.co/datasets/savalera/isear-from-original
```

For more information about specific aspects of the emotional responses in the original study, see also:

- Scherer, K. R. (1997). Profiles of emotion-antecedent appraisal: testing theoretical predictions across cultures. Cognition and Emotion, 11, 113-150.
- Scherer, K. R. (1997). The role of culture in emotion-antecedent appraisal. Journal of Personality and Social Psychology, 73, 902-922.

### Contributions

This dataset preparation, cleaning, and documentation was contributed by Savalera, making valuable psychological research data accessible for modern medical and safety applications.

## License

Apache License 2.0